Medication Administration Errors in a Mental Health Hospital
Alan Cottney, East London NHS Foundation Trust
MUSN Event, 6th June 2013
The problem…
The problem…
NPSA, Safety in Doses 2009
The problem…
NPSA, Safety in Doses 2009
The problem…
• Locally:
– Audit of last 4 years medication administration incident reports in ELFT
– Average number incident reports of incorrect administration of medication:
•141 per year
The problem…
• Is this the tip of the iceberg? :
– Suggested that for every 1 administration error detected by incident report, 300 could be detected if nurses were observed directly (Barker et al 2002)
– If this was the case in ELFT, the true administration error rate would be around 42,000 per year
The problem…
• “Medication error rates are important for gauging the scope of the problem, setting priorities for prevention strategies, and measuring the impact of those strategies.”
– Institute of Medicine, Preventing Medication Errors
The Project • Aims:
– To identify the incidence, nature, and severity of medication administration errors that are made at ELFT
– To investigate factors that contribute to errors
– To develop strategies for error reduction
• Sponsored by London Deanery and NHS London’s Simulation Technology-enhanced Learning Initiative (STeLI)
• How to investigate errors?
• “Observation is the most valid and effective method to detect and to quantify administration errors”
– Council of Europe
The Project
• A pharmacist or pharmacy technician observed the morning, lunch, evening and night medication rounds on each of the 45 wards in ELFT.
• The next slides outline their findings…
The Project
Ward specialities at East London NHS Foundation Trust
Speciality Number of inpatient wards
Adult mental health services 18
Forensic mental health services 15
Mental health care of older people services
8
Community health services 2
Child and adolescent mental health services
2
TOTAL 45
Details of medication rounds observed
Acute adult
PICU Forensic MHCOP CAMHS Community Health
OVERALL
Number of medication rounds observed 60 16 60 24 8 12 180
Number of service users observed receiving medication
495 116 429 193 5 179 1417
Number of doses of medication observed being given
1115 286 963 722 9 609 3704
Overall error details
Total number of administration errors observed
153
Medication rounds on which an error was made
69 (38%)
Average number of errors per medication round
0.9
Average number of errors per patient given medication
0.11
Average number of errors per dose given
0.04
Number of doses that need to be given for one error to occur
Acute adult
PICU Forensic MHCOP CAMHS Community Health
OVERALL
Average number of doses given before one administration error is made
21 48 21 33 n/a 23 24
Severity of errors
48
95
10
00
10
20
30
40
50
60
70
80
90
100
A: Negligible B: Minor C: Serious D: Fatal
Nu
mb
er
of
err
ors
Examples of errors
• Negligible:
– A patient was prescribed metformin liquid, but was administered metformin tablets.
• Minor
– A patient was prescribed clozapine, but this was unintentionally omitted.
Examples of errors
• Serious
– A patient had recently suffered a deep vein thrombosis, and was on the full treatment dose of a low-molecular weight heparin (LMWH). The LMWH dose was unintentionally omitted due to an oversight on the part of the administering nurse.
60
19
2
8
19 18
2
15
1 1 0
8
0
10
20
30
40
50
60
70
Dos
e om
ission
Wro
ng dos
e
Wro
ng stre
ngth/con
centratio
n
Wro
ng dru
g
Wro
ng fo
rm
Wro
ng te
chniqu
e
Wro
ng rou
te
Wro
ng time
Wro
ng patient
Mon
itorin
g er
ror
Exp
ired dr
ug
Other
Nu
mb
er
of
err
ors
Medication involved in observed administration errors
Mental health
36%
Physical health
64%
Mental health
36%
Physical health
64%
Type of medication involved in error
Type of medication
0
5
10
15
20
25
Antidep
ress
ant
Anti-inf
ectiv
e
Antim
anic
Antipsy
chot
ic
Benzo C
V
Diabe
tic
Hyo
scine
Inha
ler
Laxa
tive
Prom
etha
zine
Simple
analge
sia
Topical
Nu
mb
er
of
Err
ors
Statistically significant factors
• The following variables were found to independently predict the number of errors: – Total number of regular doses due
– Total number of PRN doses given
– Total number of other ward activities
– Trolley medication stored according in alphabetical or no order, rather than by drug class
Statistics • A Poisson regression with robust standard errors
was used to determine the best combination of predictors and the relative risks of these predictors to predict error.
• Variables which had a p-value of <= 0.2 were included in the modelling analysis.
• Predictors were included in the final model if they were significant at the 0.05 level (WALD test). All potential correlations were checked and variables were removed if they were highly correlated with other predictors included in the model.
Number of doses
• For every increment of one regular dose of medication due an error is 3% more likely to occur on that medication round
(RR:1.03: 1.02-1.03 95% CI, p= <0.0001).
Effect of PRN doses
No PRN doses given One or more PRN doses given
Number of medication rounds
93 87
Number of patients 668 749
Number of doses given
1833 1871
Number of errors made
51 102
PRN Doses
• An increase of one PRN dose given predicts a 17% increase in error per round
(RR:1.17; 1.05-1.30 95% CI, p=0.041).
Ward activities
• For every ward activity happening at the same time as the medication round, the likelihood of an error on that round increases by 33%
(RR: 1.33; 1.10-1.62 95% CI, p=0.0038).
Error rate per ward activity
Activity Number of medication rounds during which activity was also taking place
Subsequent average number of errors per dose
No activity 97 0.030
Meal 53 0.051
OT group 10 0.054
Smoking break 31 0.050
Ward meeting 15 0.053
Ward round 23 0.060
Other factors that may have influenced error rate…
Morning and night medication administration round compared
Morning round Night round
Average number of doses given per patient 3.5 2.5
Average number of interruptions per round 4.5 2.2
Average number of other activities at same time as medication round
1.3 0.2
Average number of staff members down 0.4 0.1
Effect of nursing band on error rate
Nurse AfC Band
Number of rounds
Errors per dose
Band 5 120 0.03
Band 6 44 0.06
Band 7 7 0.05
Not stated 9 0.08
Permanent v bank staff nurses
Nurse status Number of rounds
Errors per dose
Permanent 164 0.04
Bank 11 0.06
Not stated 5 0
Ward staffing
Nurses down on usual allocation
Number of rounds
Errors per dose
No staff down 141 0.04
One or more staff members down
39 0.05
Any-type interruption
R2 = 0.8942
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
No
interruptions
1 interruption 2 interruption 3 interruption 4 interruption 5 or more
interruptions
Nu
mb
er
of
err
ors
/med
icati
on
ro
un
d
Source of interruptions
Total number of interruptions
Average number of interruptions per medication
round
From patients 257 1.43
From staff members
242 1.34
Carrying out other ward activities
38 0.21
Interruptions from staff members
0
20
40
60
80
100
120
Morning Lunch Evening Night
To
tal
nu
mb
er
of
inte
rru
pti
on
s
Association between expiry date checking and administration errors
All medication expiry dates checked
Not all medication expiry dates checked
Percentage of patients
31% 69%
Errors/patient 0.074 0.110
Prescription reading • In 52% (80/153) of the administration errors the
pharmacist believed a lack of thoroughness when reading the prescription was a contributing factor. Some examples of these types of errors are: – A nurse misread the chart and administered
medication due at 18:00hrs at 22:00hrs. – Six doses of medication were unintentionally omitted
because the nurse failed to notice that the patient had a second prescription chart.
– A clozapine dose was unintentionally omitted because the nurse did not look at the titration sheet on the front of the chart.
Nurse knowledge • The pharmacy observers identified a lack of knowledge on the
part of the administering nurse as a contributing factor in 35% (53/153) of the errors that were made. Examples of these errors include: – Cutting tablets that should not be cut e.g. modified release
tablets. – Not realising importance of timeliness of certain medication
e.g. insulin, antibiotics. – The patient used their inhaler incorrectly, but this was not
corrected by the nurse because they were unaware of how it was suppose to be used.
– Nurse thought that by pulling syringe plunger to the end of a 5ml syringe this would give a volume of 5ml; not realising that the graduations stopped before the end.
– Unaware of the difference between vitamin B and multivitamin tablets.
Similar drugs • Similar drugs being mixed up was identified as the cause
of 11% of the errors. Examples include: – Venlafaxine modified release and immediate release
tablets. – Olanzapine orodispersible and normal release tablets. – Haloperidol tablets and liquid. – Valproate modified release and enteric coated tablets. – Lansoprazole orodispersible tablets and normal-release
capsules. – Aripiprazole liquid and tablets. – Soluble Adcal and chewable tablets. – Hyoscine hydrobromide and hyoscine butylbromide tablets – Carmellose and hypromellose eye drops. – Humalog and humulin insulin.
Other contributing factors • Unclear prescription:
– 5% of errors
• Agitated/demanding patient: – 10% of errors
• Ward round: – 3% of errors
• Liquid medication: – 9% of errors
Intentionally omitted doses
• Of the 443 intentionally omitted doses, only 7 of the doses should not have been omitted. – So, only 1.6% of doses that were omitted
were omitted inappropriately.
• This demonstrates that the observed nurses generally make sound clinical decisions.
Reasons for intentionally omitted doses
Patient refused
58%
Clinical reasons
18%
Patient off ward
7%
Drug unavailable
7%
Topicals omitted
until wash
10%
Documentation
• Doses not signed for: 1.5%
• Given against unsigned prescription: 0.4%
• Missing consent: 2.2%
• Allergy status not completed: 0.6%
• So, on the whole, documentation was of a good standard.
DATIX v Direct observation
• Number of incident reports during observation period = 17
• Rate of error detection: – DATIX: 0.0017 per round – Observation: 0.8500 per round
• So 500 times more errors are detected by direct observation than by incident reporting via DATIX.
Next steps…
• Nurses were invited to give their responses to the findings.
• Themes in their Reponses: – Need for better education
– Changing/abolishing the times on the drug chart
– More support when giving medication: increased staff in clinical areas, interceptor for queries.
– More pharmacist input: make informal observation more common.
Next steps…
• Improved medication administration training. – Role-play medication administration
training • Incorporates training round the most
commonly observed errors
– Educational film
• Based around harnessing the cognitive power of error .
Next steps…
• Thinking about a ward safety checklist for medication administration…
STEP AREA TO CHECK WHAT TO CONFIRM
1 Front of chart
Patient name
Allergy status
Once-off dose due?
Depot dose due?
Any additional charts?
2 Inside chart –
opened fully
For each prescription:
Drug name
Time due
Dose
Route
Form (e.g. liquid/tablet)
Additional instructions?
No medication overlooked
3 PRN side (if used)
In addition to Step 3 confirmations:
Reason for use
Last given?
Next steps…
• Introduction of electronic medicines cabinets
Questions
• Any requests for further information can be directed to:
– Alan Cottney, Project Lead Pharmacist, East London NHS Foundation Trust.
References 1. Barker KN, Flynn EA, Pepper GA, et al. Medication errors
observed in 36 health care facilities. Arch Intern Med 2002; 162: 1897-903.
2. Shane, R. Current status of administration of medicines. Am J Health-Syst Pharm 2009; 66 (Suppl 3):S42-8.
3. Council of Europe Expert Group on Safe Medication Practices. Creation of a better medication safety culture in Europe: building up safe medication practices.www.gs1-health.net/downloads/medication.safety.report.2007.pdf
4. National Patient Safety Agency. The fourth report from the Patient Safety Observatory. Safety in doses: medication safety incidents in the NHS. London: NPSA, 2009.
5. Institute of Medicine (IOM) Committee on Identifying and Preventing Medication Errors. Preventing Medication Errors. Washington, DC: The National Academies Press, 2007.